Why is ++i faster than i++ in C++?

I've heard this question come up a few times in C++ programming circles, "Why is ++i faster/better than i++?"

The short answer is: i++ has to make a copy of the object and ++i does not.

The long answer involves some code examples.

int i = 1;
int j = i++; // j is equal to 1, because i is incremented after the value is returned, 
              // which means a copy was made of the old value
int k = ++i; // k is equal to 3, because i is incremented, then returned. No copy is needed

Or a more robust example of the operator overloads:

class integer
{
public:
  integer(int t_value)
    : m_value(t_value)
  {
  }
 
  int get_value() { return m_value; }
 
  integer &operator++() // pre increment operator
  {
    ++m_value;
    return *this;
  }
 
  integer operator++(int) // post increment operator
  {
    integer old = *this; // a copy is made, it's cheap, but it's still a copy
    ++m_value;
    return old; // Return old copy
  }
 
private:
  int m_value;

You might ask, "why cannot the C++ compiler optimize away the copy of you don't use it?" Really:

int i=1;
++i; //why is this
i++; //faster than this if I'm not using the copied returned value?

It's possible that the compiler may optimize away the copy for you, but it's also possible that your post-increment operator overload has other side effects that the compiler cannot optimize for without changing the meaning of your code.

On a side note, it's rare that you literally want to increment and return the old value, so ++i is often more semantically correct than i++, besides being more efficient.

Comments

Hello, just a note : nearly

Hello,
just a note : nearly every compiler will make "i++;" standalone expression as fast as "++i;" when "i" is a variable of the built-in type.
But usually, this is not true for the more complex user-defined types (like iterators)..

Anyway, I think that by no doubt, it is a good habit to prefer "++i" construction whenever possible.